Scan barcode
A review by alykat_reads
Noise: A Flaw in Human Judgment by Cass R. Sunstein, Daniel Kahneman, Olivier Sibony
slow-paced
1.5
This could have been about 300 pages shorter. As a statistician, I do appreciate the palatable explanation of standard deviations, bell curves, medians, and what that information means when regarding a set of data and how to interpret it; as time and time again I see people just butcher the analysis of a set of data and have no idea what they're actually saying.
This book is just so redundant though, to the point where it is painful. And while it may be an interesting factoid that judges tend to dole out lighter sentences on days with good weather, there's not much else that can be done with that information. The American 'justice' system is already known to have racism intertwined, with blacks getting much harsher sentences than whites when they have the same criminal history and the crime is similar. So yes, we know it's "noisy." There are situations (e.g. the APGAR test) where noise has been reduced, but it's not possible to entirely eliminate noise. As a former underwriter, I wasn't surprised at all that there was "noise" when it came to that, but this was written as though there was nothing that could be done. There's easily parameters that can be put into place when underwriting to vastly reduce the discrepancies between the underwriters in the company.
The author also separates noise from bias, but I'm not entirely sure that we can eradicate one from the other; or that more often than not, they're the same thing.
Idk, it was just very redundant and not much more information than "people use biases to make judgments."
This book is just so redundant though, to the point where it is painful. And while it may be an interesting factoid that judges tend to dole out lighter sentences on days with good weather, there's not much else that can be done with that information. The American 'justice' system is already known to have racism intertwined, with blacks getting much harsher sentences than whites when they have the same criminal history and the crime is similar. So yes, we know it's "noisy." There are situations (e.g. the APGAR test) where noise has been reduced, but it's not possible to entirely eliminate noise. As a former underwriter, I wasn't surprised at all that there was "noise" when it came to that, but this was written as though there was nothing that could be done. There's easily parameters that can be put into place when underwriting to vastly reduce the discrepancies between the underwriters in the company.
The author also separates noise from bias, but I'm not entirely sure that we can eradicate one from the other; or that more often than not, they're the same thing.
Idk, it was just very redundant and not much more information than "people use biases to make judgments."